Changes which Maarit has already put into the PDF are marked with ***. Comments for internal discussion/action in []. General aspects like wrong footnote numbering or wrong symbols for mean EMF and force are not mentioned here. +++ Added to the pdf XXX not yet added XXX Sec. 2.3.3, l. 6/7: "50 time units" -> "$0.5 viscous times$" XXX p. 4, Sec. 2.6, l. 3: insert behind [new] ${\cal A}/k_1$: ", further the viscous time scale $T_\nu=(\nu k_1^2)^{-1}$." XXX p. 5, Table 1, caption: [new] "S=-25 \nu k_1^2" -> "S=-25/T_\nu" end of 1st para: add "In FMHD, the sound speed was set to $100/k_1 T_\nu$." XXX p. 6, left col., l. 5 from below: delete ", with equal contributions from the kinetic and magnetic forcings" XXX p. 7, Sec. 3.2.1, l.4: [new] "S=-25 \nu k_1^2" -> "S=-25/T_\nu" XXX p. 9, Sec. 3.2.3, l.4: [new] "S=-25 \nu k_1^2" -> "S=-25/T_\nu" XXX p. 15, left col., l. 9: "50 (in code units)" -> "$0.5 T_\nu$" p. 8, left col.. [l. 2: "exceeds the corresponding fluctuations in α" [How would we compare?] %MJK We discussed this already before. It IS a bit questionable to compare magnitudes %MJK of alphas to etas, but now we at least compare normalized values, which should bear %MJK some significance. The point here is that alpha fluctuations should be zero, and %MJK hence mark the "noise" level, and that is why it is important to say that the eta %MJK values exceed the noise level. We do have the error bars, of course, but I would %MJK still think it is good to mention this. %MR: Can't we compare \alpha and \eta k_z? %MJK This we also discussed in length [the normalization of alpha_rms and eta_rms in the tables]. %MJK Now it is far too late to change this. %MJK We have more rigorous comparison with the 0D model then, and most, if not all, conclusions are %MJK anyways based on those. %MJK I would leave it. %MR: A proper comparison would be between \alpha/(\kf \eta_0) and \eta/eta_0 (k_z/\kf). So \eta/eta_0 has to be damped by a factor \le 5. % I think we should therefore leave out the sentence "The magnitude of the normalized η yx values, however, exceeds the corresponding fluctuations in α.". [l. 6: "α values are similar or a bit larger, and clearly exceed the η_yx component" How to compare?] %MJK See my comment above. Now important to say from the perspective that eta_yx is very small. %MJK See my comment above. %MR: The same here. XXX p. 9, right col, l. 3: "the wavenumber" -> "$k_z/k_{1z}$" %MJK I think it is fine without this change. Also, in the table, we have now 2\pi, and this seems to be conflicting with Matthias changes. Matthias, please check throughout. %MR: The table is fine now, but saying "the wavenumber" would implicitly refer to our length scale and k_z in terms of k_1 is actually hardly growing. +++p. 11, left col. above (26): "effect is" -> "effect through $\eta_{yx}$," XXX right col., Eq.(27): "\partial_t \bm{A}" [new] -> "\partial_t \overline{\bm{A}}" XXX (28): make outer brackets bigger XXX 2nd l. below Eq. (28): add behind "zero means," "and standard deviations equal to the respective rms values," +++ 2nd para, l. 7: magnetic -> magneto-kinetic +++ l. 3 from below: cause -> causes +++ 3rd para, l. 6 from below: add "in Table 3" behind "with an asterisk" +++p. 12: Fig 6, caption end: k^2 -> k_z^2 +++ left col., l. 3: "SC" -> "SC effect" +++ l. 7: "$D_{\eta S}$" -> "$D_{\eta_{\rm rms} S}$" *** +++ l. 7 from below: "SC" -> "SC-effect" +++ right col. last l.: kinetic and magnetic -> magneto-kinetic +++p. 13, left col., l. 4: "damp" -> "damp down" +++ l. 6: dynamo numbers -> $D_{\alpha S}$ +++ last para, l. 1: magnetic -> magneto-kinetic XXX l. 7 from below: no comma [not properly showing up] +++ right col l. 7: "such that it decreases" -> "such that its modulus decreases" p. 14: right col., [l. 3: m=2 self explaining?] %MJK I would think so. Make a suggestion how to improve! %MR: A suggestion is not so easy. That's why I asked. If we agree that it is self-explaining we leave it as it is. %MJK We leave it then? %MR: Perhaps we ask Axel. +++p. 15: Table 4: don't inflate to full width [Run SKM1a007 not explained] %MJK It is, indeed, missing from Table 2. I wonder why. This would be a question to Javier, but we will not get an answer, unfortunately. %MJK I will try to search for the data, but I think it is bad to start adding entries to the data table at proofs. We should maybe %MJK explain it off otherwise. It is obviously a run very close to 008 and 006 that are in the table. %MJK Found the data and could compute the numbers with the 'ju' method SKM1a007 & $-$0.649 &2.077$\pm$0.025 &2.051$\pm$0.007 &0.120$\pm$0.018 &$-$0.177$\pm$0.024 &0.091$\pm$0.027 &0.173$\pm$0.068 \\%SMK1a007 in Puhti %MJK Do we want to add this or not at proofs? %MR: Perhaps not in a Table, but somehow describing the run would be good. %MJK It is not included in the plots, hence I would leave it out from the table... I cannot remember why it was not included in the plots... %MJK irritating... Maybe I could still add it to Fig. 4... I will keep thinking about this... %MR: The following not needed if we put SKM1a007 into Table 2. XXX p. 15, left col., l. 2: "Run SKM1a007" -> "the representative Run SKM1a007" XXX Table 4: add "Notes: Standard forcing with $\kf/k_1=5$, $\ShK=-0.649$, $\Pm=1$, ${\cal A} = 1$, $\Rm = ...$." +++ Fig. 10, caption: add "Note that the {\sf jb} results are almost completely on top of the {\sf bu} ones." App. 4, l. 4: "and assume" -> ", assume" Eq. (C1): sizes of left and right bracket pairs () and < > don't match more space between \tau and I (2 times) put (C1) on first line p. 16, left col., (C3): too long, bracket sizes don't match l. 1: y_0 -> y l. 3: "\rho_0 \mu_0" -> "\rho"; [...] -> (...) l. 4: first 0 should be bold (C5)-(C6): align first + and - on each line number each line because of that: below (C7): "the second contribution in (C5)" -> "(C6)" "by parts in (C6)" -> "by parts in (C7)" right col.: "(C6) and (C7)" -> "(C7) -- (C9)" "the first term in (C5)" -> "(C5)" right col., in 2.: insert "-effect" behind MSC no indentation before (